Cluster, the term I have run across for so many times while reading some technical articles, and I tried googling detailed explanations for it but found few that really satisfied me. though, after reading those explanations, I've gained a shallow understanding of what it is about. I know that clustering basically means putting serveral servers(here I am refering to Web servers like Resin, Tomcat, HTTP Servers like Apache...) together and letting them all fulfil requests from the client to condense to processing time, and I think that there should be more features that clustering has to provide, so what's that?
And the most important, I want to know whether clustering will keep sessions(HttpSession in the Servlet API) for other nodes(servers) in the cluster so they can share all these sessions to remember the users that have logged in. if so, how is that achieved?
Our company, maintaining a wapsite(site that viewd via mobile phones), has implemented a distributed environment by persisting all sessions into database for supporting session sharing across different servers.(ps: a wap site can't keep HttpSession for mobile phone users, cause most of the mobile phones don't support cookie, and the request strings of the URL sent by mobile phones would always be filtered by the telecommunication company), for a wap site, persisting sessions into database is a one and only option(am I right?), and every server has to read the session data from database for its own use, now the issue is how the servers can be more efficient in sharing the session data? is it possible for all servers to share the session data without fetching them from database every time the session data are required in the appllication? can CLUSTERING address this issue?
Your help will be GREATLY, GREATELY appreciated!!! Thank you in advance. --Kevin Tse